pacman::p_load(olsrr, ggstatsplot, ggpubr, sf, spdep, GWmodel, tmap, tidyverse, gtsummary, performance, see, sfdep)In-class Ex 7
1. Overview
In this in-class exercise, we build a hedonic pricing model for condominiums using Geographically Weighted Regression (GWR) to account for spatial non-stationarity, where relationships between variables differ across geographic locations. This exercise builds on hands-on exercise 7, with additional emphasis on other R packages relevant to GWR
2. Importing & Transforming Data
We will import and work with two datasets:
- URA’s 2014 master plan subzone boundary
- Condominium resale prices 2015
We use st_transform() to assign the correct ESPG code 3414.
mpsz <- st_read(dsn = "data/geospatial", layer = "MP14_SUBZONE_WEB_PL") %>%
st_transform(3414)Reading layer `MP14_SUBZONE_WEB_PL' from data source
`/Users/stephentay/stephentay/ISSS626-Geospatial-Analytics/In-class_Ex/In-class_Ex07/data/geospatial'
using driver `ESRI Shapefile'
Simple feature collection with 323 features and 15 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 2667.538 ymin: 15748.72 xmax: 56396.44 ymax: 50256.33
Projected CRS: SVY21
We import the condo resale prices dataset using read_csv().
condo_resale = read_csv("data/aspatial/Condo_resale_2015.csv")
head(condo_resale)# A tibble: 6 × 23
LATITUDE LONGITUDE POSTCODE SELLING_PRICE AREA_SQM AGE PROX_CBD
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1.29 104. 118635 3000000 309 30 7.94
2 1.33 104. 288420 3880000 290 32 6.61
3 1.31 104. 267833 3325000 248 33 6.90
4 1.31 104. 258380 4250000 127 7 4.04
5 1.32 104. 467169 1400000 145 28 11.8
6 1.31 104. 466472 1320000 139 22 10.3
# ℹ 16 more variables: PROX_CHILDCARE <dbl>, PROX_ELDERLYCARE <dbl>,
# PROX_URA_GROWTH_AREA <dbl>, PROX_HAWKER_MARKET <dbl>,
# PROX_KINDERGARTEN <dbl>, PROX_MRT <dbl>, PROX_PARK <dbl>,
# PROX_PRIMARY_SCH <dbl>, PROX_TOP_PRIMARY_SCH <dbl>,
# PROX_SHOPPING_MALL <dbl>, PROX_SUPERMARKET <dbl>, PROX_BUS_STOP <dbl>,
# NO_Of_UNITS <dbl>, FAMILY_FRIENDLY <dbl>, FREEHOLD <dbl>,
# LEASEHOLD_99YR <dbl>
Since the CSV file contains latitude and longitude coordinates, but the coordinate system is unknown, we make an informed guess and assign the closest geographic coordinate system—EPSG 4326. (You could verify the accuracy by plotting the data to ensure the spatial points align with real-world locations). If confirmed, we transform the data to the projected coordinate system EPSG 3414.
condo_resale_sf = condo_resale %>%
st_as_sf(coords = c("LONGITUDE", "LATITUDE"),
crs=4326) %>% # use the geo coordinate system of the lat/long in the csv file
st_transform(crs=3414)
head(condo_resale_sf)Simple feature collection with 6 features and 21 fields
Geometry type: POINT
Dimension: XY
Bounding box: xmin: 22085.12 ymin: 29951.54 xmax: 41042.56 ymax: 34546.2
Projected CRS: SVY21 / Singapore TM
# A tibble: 6 × 22
POSTCODE SELLING_PRICE AREA_SQM AGE PROX_CBD PROX_CHILDCARE PROX_ELDERLYCARE
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 118635 3000000 309 30 7.94 0.166 2.52
2 288420 3880000 290 32 6.61 0.280 1.93
3 267833 3325000 248 33 6.90 0.429 0.502
4 258380 4250000 127 7 4.04 0.395 1.99
5 467169 1400000 145 28 11.8 0.119 1.12
6 466472 1320000 139 22 10.3 0.125 0.789
# ℹ 15 more variables: PROX_URA_GROWTH_AREA <dbl>, PROX_HAWKER_MARKET <dbl>,
# PROX_KINDERGARTEN <dbl>, PROX_MRT <dbl>, PROX_PARK <dbl>,
# PROX_PRIMARY_SCH <dbl>, PROX_TOP_PRIMARY_SCH <dbl>,
# PROX_SHOPPING_MALL <dbl>, PROX_SUPERMARKET <dbl>, PROX_BUS_STOP <dbl>,
# NO_Of_UNITS <dbl>, FAMILY_FRIENDLY <dbl>, FREEHOLD <dbl>,
# LEASEHOLD_99YR <dbl>, geometry <POINT [m]>
3. Building the Hedonic Price Model
The following steps are taken to build the hedonic price model.
3.1 Correlation Analysis
To prevent multicollinearity, it is essential to examine relationships between variables to identify those with high correlations. While immediate removal isn’t necessary, these variables should be closely monitored during the multicollinearity test later.
As an alternative to the corrplot package, the ggcorrmat() function from the ggstatsplot package can also be used for correlation analysis.
ggcorrmat(condo_resale[, 5:23])
3.2 Initial Hedonic Pricing Model Using MLR
We begin by building an initial MLR model with all variables deemed relevant for predicting the selling price. We will assess the model in the next step.
condo_mlr <- lm(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE + PROX_ELDERLYCARE +
PROX_URA_GROWTH_AREA + PROX_HAWKER_MARKET + PROX_KINDERGARTEN +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_TOP_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_SUPERMARKET +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD +
LEASEHOLD_99YR,
data=condo_resale_sf)
summary(condo_mlr)
Call:
lm(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD + PROX_CHILDCARE +
PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA + PROX_HAWKER_MARKET +
PROX_KINDERGARTEN + PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_TOP_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_SUPERMARKET +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD +
LEASEHOLD_99YR, data = condo_resale_sf)
Residuals:
Min 1Q Median 3Q Max
-3471036 -286903 -22426 239412 12254549
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 543071.4 136210.9 3.987 7.03e-05 ***
AREA_SQM 12688.7 370.1 34.283 < 2e-16 ***
AGE -24566.0 2766.0 -8.881 < 2e-16 ***
PROX_CBD -78122.0 6791.4 -11.503 < 2e-16 ***
PROX_CHILDCARE -333219.0 111020.3 -3.001 0.002734 **
PROX_ELDERLYCARE 170950.0 42110.8 4.060 5.19e-05 ***
PROX_URA_GROWTH_AREA 38507.6 12523.7 3.075 0.002147 **
PROX_HAWKER_MARKET 23801.2 29299.9 0.812 0.416739
PROX_KINDERGARTEN 144098.0 82738.7 1.742 0.081795 .
PROX_MRT -322775.9 58528.1 -5.515 4.14e-08 ***
PROX_PARK 564487.9 66563.0 8.481 < 2e-16 ***
PROX_PRIMARY_SCH 186170.5 65515.2 2.842 0.004553 **
PROX_TOP_PRIMARY_SCH -477.1 20598.0 -0.023 0.981525
PROX_SHOPPING_MALL -207721.5 42855.5 -4.847 1.39e-06 ***
PROX_SUPERMARKET -48074.7 77145.3 -0.623 0.533273
PROX_BUS_STOP 675755.0 138552.0 4.877 1.20e-06 ***
NO_Of_UNITS -216.2 90.3 -2.394 0.016797 *
FAMILY_FRIENDLY 142128.3 47055.1 3.020 0.002569 **
FREEHOLD 300646.5 77296.5 3.890 0.000105 ***
LEASEHOLD_99YR -77137.4 77570.9 -0.994 0.320192
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 755800 on 1416 degrees of freedom
Multiple R-squared: 0.652, Adjusted R-squared: 0.6474
F-statistic: 139.6 on 19 and 1416 DF, p-value: < 2.2e-16
3.3 Model Assessment (using olsrr)
In model assessment, we evaluate the model’s p-value and R² to determine overall significance and explanatory power. Next, we assess individual variables to identify any that are non-significant, as these should be removed to enhance the model’s robustness.
The olsrr package is used to generate a comprehensive, tidy report for assessing the MLR model. The summary report indicates that the model is statistically significant (p < 0.05) and explains 64.7% of the variance in the dependent variable. However, since some independent variables are not statistically significant, they should be excluded from the final model.
ols_regress(condo_mlr) Model Summary
-----------------------------------------------------------------------------
R 0.807 RMSE 750537.537
R-Squared 0.652 MSE 571262902261.220
Adj. R-Squared 0.647 Coef. Var 43.160
Pred R-Squared 0.637 AIC 42971.173
MAE 412117.987 SBC 43081.835
-----------------------------------------------------------------------------
RMSE: Root Mean Square Error
MSE: Mean Square Error
MAE: Mean Absolute Error
AIC: Akaike Information Criteria
SBC: Schwarz Bayesian Criteria
ANOVA
--------------------------------------------------------------------------------
Sum of
Squares DF Mean Square F Sig.
--------------------------------------------------------------------------------
Regression 1.515738e+15 19 7.977571e+13 139.648 0.0000
Residual 8.089083e+14 1416 571262902261.220
Total 2.324647e+15 1435
--------------------------------------------------------------------------------
Parameter Estimates
-----------------------------------------------------------------------------------------------------------------
model Beta Std. Error Std. Beta t Sig lower upper
-----------------------------------------------------------------------------------------------------------------
(Intercept) 543071.420 136210.918 3.987 0.000 275874.535 810268.305
AREA_SQM 12688.669 370.119 0.579 34.283 0.000 11962.627 13414.710
AGE -24566.001 2766.041 -0.166 -8.881 0.000 -29991.980 -19140.022
PROX_CBD -78121.985 6791.377 -0.267 -11.503 0.000 -91444.227 -64799.744
PROX_CHILDCARE -333219.036 111020.303 -0.087 -3.001 0.003 -551000.984 -115437.089
PROX_ELDERLYCARE 170949.961 42110.748 0.083 4.060 0.000 88343.803 253556.120
PROX_URA_GROWTH_AREA 38507.622 12523.661 0.059 3.075 0.002 13940.700 63074.545
PROX_HAWKER_MARKET 23801.197 29299.923 0.019 0.812 0.417 -33674.725 81277.120
PROX_KINDERGARTEN 144097.972 82738.669 0.030 1.742 0.082 -18205.570 306401.514
PROX_MRT -322775.874 58528.079 -0.123 -5.515 0.000 -437586.937 -207964.811
PROX_PARK 564487.876 66563.011 0.148 8.481 0.000 433915.162 695060.590
PROX_PRIMARY_SCH 186170.524 65515.193 0.072 2.842 0.005 57653.253 314687.795
PROX_TOP_PRIMARY_SCH -477.073 20597.972 -0.001 -0.023 0.982 -40882.894 39928.747
PROX_SHOPPING_MALL -207721.520 42855.500 -0.109 -4.847 0.000 -291788.613 -123654.427
PROX_SUPERMARKET -48074.679 77145.257 -0.012 -0.623 0.533 -199405.956 103256.599
PROX_BUS_STOP 675755.044 138551.991 0.133 4.877 0.000 403965.817 947544.272
NO_Of_UNITS -216.180 90.302 -0.046 -2.394 0.017 -393.320 -39.040
FAMILY_FRIENDLY 142128.272 47055.082 0.056 3.020 0.003 49823.107 234433.438
FREEHOLD 300646.543 77296.529 0.117 3.890 0.000 149018.525 452274.561
LEASEHOLD_99YR -77137.375 77570.869 -0.030 -0.994 0.320 -229303.551 75028.801
-----------------------------------------------------------------------------------------------------------------
3.4 Checking Multicollinearity
We use the following code to compute the Variance Inflation Factor (VIF) to help us identify multicollinearity. A VIF between 5 and 10 indicates moderate multicollinearity and requires monitoring, while a VIF above 10 suggests severe multicollinearity, warranting variable elimination.
As all VIF values are below 10, no variables need to be removed.
ols_vif_tol(condo_mlr) Variables Tolerance VIF
1 AREA_SQM 0.8601326 1.162611
2 AGE 0.7011585 1.426211
3 PROX_CBD 0.4575471 2.185567
4 PROX_CHILDCARE 0.2898233 3.450378
5 PROX_ELDERLYCARE 0.5922238 1.688551
6 PROX_URA_GROWTH_AREA 0.6614081 1.511926
7 PROX_HAWKER_MARKET 0.4373874 2.286303
8 PROX_KINDERGARTEN 0.8356793 1.196631
9 PROX_MRT 0.4949877 2.020252
10 PROX_PARK 0.8015728 1.247547
11 PROX_PRIMARY_SCH 0.3823248 2.615577
12 PROX_TOP_PRIMARY_SCH 0.4878620 2.049760
13 PROX_SHOPPING_MALL 0.4903052 2.039546
14 PROX_SUPERMARKET 0.6142127 1.628100
15 PROX_BUS_STOP 0.3311024 3.020213
16 NO_Of_UNITS 0.6543336 1.528272
17 FAMILY_FRIENDLY 0.7191719 1.390488
18 FREEHOLD 0.2728521 3.664990
19 LEASEHOLD_99YR 0.2645988 3.779307
3.5 Variable Selection
We use the ols_step_forward_p() function to perform stepwise forward selection. Although there are other criteria that could guide the selection process, we prioritise the p-value to ensure that all variables included in the final model are statistically significant.
We can visualise the stepwise forward selection process using the plot() function, which displays the incremental changes in Adjusted R², AIC, and RMSE throughout the selection process.
condo_fw_mlr <- ols_step_forward_p(condo_mlr,
p_val = 0.05,
details = FALSE)plot(condo_fw_mlr)
3.6 Visualising Model Parameters
The following method allows us to visualise all the model parameters.
ggcoefstats(condo_mlr, sort = "ascending")
3.7 Testing for Non-linearity
It is important to test the assumption of linearity and additivity in the relationship between the dependent and independent variables. The figure shows that most data points are scattered around the zero line, indicating that the relationships between the dependent and independent variables are linear.
ols_plot_resid_fit(condo_fw_mlr$model)
3.8 Testing Normality of Residuals
We use ols_plot_resid_hist() and ols_test_normality() to check the normality of the residuals.
ols_plot_resid_hist(condo_fw_mlr$model)
ols_test_normality(condo_fw_mlr$model)-----------------------------------------------
Test Statistic pvalue
-----------------------------------------------
Shapiro-Wilk 0.6856 0.0000
Kolmogorov-Smirnov 0.1366 0.0000
Cramer-von Mises 121.0768 0.0000
Anderson-Darling 67.9551 0.0000
-----------------------------------------------
3.9 Testing for Spatial Autocorrelation
The hedonic model incorporates geographically referenced attributes, making it essential to visualize the residuals. To test for spatial autocorrelation, we must convert the condo_resale.sf dataset from an sf object to a SpatialPointsDataFrame.
First, we export the residuals from the hedonic pricing model and save them as a data frame.
mlr_output <- as.data.frame(condo_fw_mlr$model$residuals) %>%
rename(`FW_MLR_RES` = `condo_fw_mlr$model$residuals`)Next, we will join the newly created dataframe with condo_resale_sf object.
condo_resale_sf <- cbind(condo_resale_sf,
mlr_output$FW_MLR_RES) %>%
rename(`MLR_RES` = `mlr_output.FW_MLR_RES`)3.10 Map Plot of Residuals
We plot the residuals on a map to identify areas of overestimation and underestimation. Visible clusters of over- or under-estimated prices may indicate the presence of spatial autocorrelation.
tmap_mode("view")
#tmap_options(check.and.fix = TRUE) -- add this code here to fix any layers with problematic lines/polygons.
tm_shape(mpsz) +
tmap_options(check.and.fix = TRUE) + # add this line here to explicitly fix problematic polygons in this specific layer.
tm_polygons(alpha = 0.4) +
tm_shape(condo_resale_sf) +
tm_dots(col = "MLR_RES",
alpha = 0.6,
style="quantile") +
tm_view(set.zoom.limits = c(11,14))tmap_mode("plot")3.11 Spatial Stationarity Test
We compute the distance-based weight matrix using sfdep package.
condo_resale_sf <- condo_resale_sf %>%
mutate(nb = st_knn(geometry, k = 6, longlat = FALSE),
wt = st_weights(nb, style = "W"),
.before = 1)To confirm the presence of spatial autocorrelation, we perform the Moran’s I test.
- H₀: The residuals are randomly distributed (spatially stationary).
- H₁: The residuals exhibit spatial dependence (spatially non-stationary).
We conduct a Global Moran’s I permutation test to determine whether spatial autocorrelation exists in the residuals.
The Global Moran’s I test for residual spatial autocorrelation shows that it’s p-value is less than the alpha value of 0.05. Hence, we will reject the null hypothesis that the residuals are randomly distributed. Since the Observed Global Moran I = 0.25586 which is greater than 0, we can infer than the residuals resemble cluster distribution.
global_moran_perm(condo_resale_sf$MLR_RES,
condo_resale_sf$nb,
condo_resale_sf$wt,
alternative = "two.sided",
nsim = 99)
Monte-Carlo simulation of Moran I
data: x
weights: listw
number of simulations + 1: 100
statistic = 0.32254, observed rank = 100, p-value < 2.2e-16
alternative hypothesis: two.sided
4. Building GWR Model
In this section, we will build the hedonic pricing models using GWR.
4.1 Fixed Bandwidth GWR Model
In the code below, the bw.gwr() function from the GWmodel package is used to determine the optimal fixed bandwidth for the model. Setting the adaptive argument to FALSE specifies that a fixed bandwidth will be used.
There are two methods to define the stopping rule: the cross-validation (CV) approach and the AIC corrected (AICc) approach. We define the stopping rule using approach agreement.
bw_fixed <- bw.gwr(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE +
PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale_sf,
approach="CV",
kernel = "gaussian",
adaptive = FALSE,
longlat = FALSE)Fixed bandwidth: 17660.96 CV score: 8.259118e+14
Fixed bandwidth: 10917.26 CV score: 7.970454e+14
Fixed bandwidth: 6749.419 CV score: 7.273273e+14
Fixed bandwidth: 4173.553 CV score: 6.300006e+14
Fixed bandwidth: 2581.58 CV score: 5.404958e+14
Fixed bandwidth: 1597.687 CV score: 4.857515e+14
Fixed bandwidth: 989.6077 CV score: 4.722431e+14
Fixed bandwidth: 613.7939 CV score: 1.379526e+16
Fixed bandwidth: 1221.873 CV score: 4.778717e+14
Fixed bandwidth: 846.0596 CV score: 4.791629e+14
Fixed bandwidth: 1078.325 CV score: 4.751406e+14
Fixed bandwidth: 934.7772 CV score: 4.72518e+14
Fixed bandwidth: 1023.495 CV score: 4.730305e+14
Fixed bandwidth: 968.6643 CV score: 4.721317e+14
Fixed bandwidth: 955.7206 CV score: 4.722072e+14
Fixed bandwidth: 976.6639 CV score: 4.721387e+14
Fixed bandwidth: 963.7202 CV score: 4.721484e+14
Fixed bandwidth: 971.7199 CV score: 4.721293e+14
Fixed bandwidth: 973.6083 CV score: 4.721309e+14
Fixed bandwidth: 970.5527 CV score: 4.721295e+14
Fixed bandwidth: 972.4412 CV score: 4.721296e+14
Fixed bandwidth: 971.2741 CV score: 4.721292e+14
Fixed bandwidth: 970.9985 CV score: 4.721293e+14
Fixed bandwidth: 971.4443 CV score: 4.721292e+14
Fixed bandwidth: 971.5496 CV score: 4.721293e+14
Fixed bandwidth: 971.3793 CV score: 4.721292e+14
Fixed bandwidth: 971.3391 CV score: 4.721292e+14
Fixed bandwidth: 971.3143 CV score: 4.721292e+14
Fixed bandwidth: 971.3545 CV score: 4.721292e+14
Fixed bandwidth: 971.3296 CV score: 4.721292e+14
Fixed bandwidth: 971.345 CV score: 4.721292e+14
Fixed bandwidth: 971.3355 CV score: 4.721292e+14
Fixed bandwidth: 971.3413 CV score: 4.721292e+14
Fixed bandwidth: 971.3377 CV score: 4.721292e+14
Fixed bandwidth: 971.34 CV score: 4.721292e+14
Fixed bandwidth: 971.3405 CV score: 4.721292e+14
Fixed bandwidth: 971.3396 CV score: 4.721292e+14
Fixed bandwidth: 971.3402 CV score: 4.721292e+14
Fixed bandwidth: 971.3398 CV score: 4.721292e+14
Fixed bandwidth: 971.34 CV score: 4.721292e+14
Fixed bandwidth: 971.3399 CV score: 4.721292e+14
Fixed bandwidth: 971.34 CV score: 4.721292e+14
We use the code chunk below to build the fixed bandwidth GWR model. There are observable improvements in R2 and the AICc. (Note: AICc is robust for small dataset.)
gwr_fixed <- gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE +
PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale_sf,
bw=bw_fixed,
kernel = 'gaussian',
longlat = FALSE)
gwr_fixed ***********************************************************************
* Package GWmodel *
***********************************************************************
Program starts at: 2024-10-16 23:56:50.289106
Call:
gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD +
PROX_CHILDCARE + PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH + PROX_SHOPPING_MALL +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data = condo_resale_sf, bw = bw_fixed, kernel = "gaussian",
longlat = FALSE)
Dependent (y) variable: SELLING_PRICE
Independent variables: AREA_SQM AGE PROX_CBD PROX_CHILDCARE PROX_ELDERLYCARE PROX_URA_GROWTH_AREA PROX_MRT PROX_PARK PROX_PRIMARY_SCH PROX_SHOPPING_MALL PROX_BUS_STOP NO_Of_UNITS FAMILY_FRIENDLY FREEHOLD
Number of data points: 1436
***********************************************************************
* Results of Global Regression *
***********************************************************************
Call:
lm(formula = formula, data = data)
Residuals:
Min 1Q Median 3Q Max
-3470778 -298119 -23481 248917 12234210
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 527633.22 108183.22 4.877 1.20e-06 ***
AREA_SQM 12777.52 367.48 34.771 < 2e-16 ***
AGE -24687.74 2754.84 -8.962 < 2e-16 ***
PROX_CBD -77131.32 5763.12 -13.384 < 2e-16 ***
PROX_CHILDCARE -318472.75 107959.51 -2.950 0.003231 **
PROX_ELDERLYCARE 185575.62 39901.86 4.651 3.61e-06 ***
PROX_URA_GROWTH_AREA 39163.25 11754.83 3.332 0.000885 ***
PROX_MRT -294745.11 56916.37 -5.179 2.56e-07 ***
PROX_PARK 570504.81 65507.03 8.709 < 2e-16 ***
PROX_PRIMARY_SCH 159856.14 60234.60 2.654 0.008046 **
PROX_SHOPPING_MALL -220947.25 36561.83 -6.043 1.93e-09 ***
PROX_BUS_STOP 682482.22 134513.24 5.074 4.42e-07 ***
NO_Of_UNITS -245.48 87.95 -2.791 0.005321 **
FAMILY_FRIENDLY 146307.58 46893.02 3.120 0.001845 **
FREEHOLD 350599.81 48506.48 7.228 7.98e-13 ***
---Significance stars
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 756000 on 1421 degrees of freedom
Multiple R-squared: 0.6507
Adjusted R-squared: 0.6472
F-statistic: 189.1 on 14 and 1421 DF, p-value: < 2.2e-16
***Extra Diagnostic information
Residual sum of squares: 8.120609e+14
Sigma(hat): 752522.9
AIC: 42966.76
AICc: 42967.14
BIC: 41731.39
***********************************************************************
* Results of Geographically Weighted Regression *
***********************************************************************
*********************Model calibration information*********************
Kernel function: gaussian
Fixed bandwidth: 971.34
Regression points: the same locations as observations are used.
Distance metric: Euclidean distance metric is used.
****************Summary of GWR coefficient estimates:******************
Min. 1st Qu. Median 3rd Qu.
Intercept -3.5988e+07 -5.1998e+05 7.6780e+05 1.7412e+06
AREA_SQM 1.0003e+03 5.2758e+03 7.4740e+03 1.2301e+04
AGE -1.3475e+05 -2.0813e+04 -8.6260e+03 -3.7784e+03
PROX_CBD -7.7047e+07 -2.3608e+05 -8.3599e+04 3.4646e+04
PROX_CHILDCARE -6.0097e+06 -3.3667e+05 -9.7426e+04 2.9007e+05
PROX_ELDERLYCARE -3.5001e+06 -1.5970e+05 3.1970e+04 1.9577e+05
PROX_URA_GROWTH_AREA -3.0170e+06 -8.2013e+04 7.0749e+04 2.2612e+05
PROX_MRT -3.5282e+06 -6.5836e+05 -1.8833e+05 3.6922e+04
PROX_PARK -1.2062e+06 -2.1732e+05 3.5383e+04 4.1335e+05
PROX_PRIMARY_SCH -2.2695e+07 -1.7066e+05 4.8472e+04 5.1555e+05
PROX_SHOPPING_MALL -7.2585e+06 -1.6684e+05 -1.0517e+04 1.5923e+05
PROX_BUS_STOP -1.4676e+06 -4.5207e+04 3.7601e+05 1.1664e+06
NO_Of_UNITS -1.3170e+03 -2.4822e+02 -3.0846e+01 2.5496e+02
FAMILY_FRIENDLY -2.2749e+06 -1.1140e+05 7.6214e+03 1.6107e+05
FREEHOLD -9.2067e+06 3.8074e+04 1.5169e+05 3.7528e+05
Max.
Intercept 112794435
AREA_SQM 21575
AGE 434203
PROX_CBD 2704604
PROX_CHILDCARE 1654086
PROX_ELDERLYCARE 38867861
PROX_URA_GROWTH_AREA 78515805
PROX_MRT 3124325
PROX_PARK 18122439
PROX_PRIMARY_SCH 4637517
PROX_SHOPPING_MALL 1529953
PROX_BUS_STOP 11342209
NO_Of_UNITS 12907
FAMILY_FRIENDLY 1720745
FREEHOLD 6073642
************************Diagnostic information*************************
Number of data points: 1436
Effective number of parameters (2trace(S) - trace(S'S)): 438.3807
Effective degrees of freedom (n-2trace(S) + trace(S'S)): 997.6193
AICc (GWR book, Fotheringham, et al. 2002, p. 61, eq 2.33): 42263.61
AIC (GWR book, Fotheringham, et al. 2002,GWR p. 96, eq. 4.22): 41632.36
BIC (GWR book, Fotheringham, et al. 2002,GWR p. 61, eq. 2.34): 42515.71
Residual sum of squares: 2.534069e+14
R-square value: 0.8909912
Adjusted R-square value: 0.8430418
***********************************************************************
Program stops at: 2024-10-16 23:56:50.824826
4.2 Adaptive Bandwidth GWR Model
The following code is similar to the one used for computing fixed bandwidth, except that the adaptive argument is set to TRUE. The recommended number of data points to use is 30.
bw_adaptive <- bw.gwr(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE +
PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale_sf,
approach="CV",
kernel = "gaussian",
adaptive = TRUE,
longlat = FALSE)Adaptive bandwidth: 895 CV score: 7.952401e+14
Adaptive bandwidth: 561 CV score: 7.667364e+14
Adaptive bandwidth: 354 CV score: 6.953454e+14
Adaptive bandwidth: 226 CV score: 6.15223e+14
Adaptive bandwidth: 147 CV score: 5.674373e+14
Adaptive bandwidth: 98 CV score: 5.426745e+14
Adaptive bandwidth: 68 CV score: 5.168117e+14
Adaptive bandwidth: 49 CV score: 4.859631e+14
Adaptive bandwidth: 37 CV score: 4.646518e+14
Adaptive bandwidth: 30 CV score: 4.422088e+14
Adaptive bandwidth: 25 CV score: 4.430816e+14
Adaptive bandwidth: 32 CV score: 4.505602e+14
Adaptive bandwidth: 27 CV score: 4.462172e+14
Adaptive bandwidth: 30 CV score: 4.422088e+14
We use the code chunk below to build the adaptive bandwidth GWR model. The report shows that the AICc the adaptive distance gwr is 41982.22 which is even smaller than the AICc of the fixed distance gwr of 42263.61.
gwr_adaptive <- gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE +
PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale_sf,
bw=bw_adaptive,
kernel = 'gaussian',
adaptive=TRUE,
longlat = FALSE)
gwr_adaptive ***********************************************************************
* Package GWmodel *
***********************************************************************
Program starts at: 2024-10-16 23:56:55.416896
Call:
gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD +
PROX_CHILDCARE + PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH + PROX_SHOPPING_MALL +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data = condo_resale_sf, bw = bw_adaptive, kernel = "gaussian",
adaptive = TRUE, longlat = FALSE)
Dependent (y) variable: SELLING_PRICE
Independent variables: AREA_SQM AGE PROX_CBD PROX_CHILDCARE PROX_ELDERLYCARE PROX_URA_GROWTH_AREA PROX_MRT PROX_PARK PROX_PRIMARY_SCH PROX_SHOPPING_MALL PROX_BUS_STOP NO_Of_UNITS FAMILY_FRIENDLY FREEHOLD
Number of data points: 1436
***********************************************************************
* Results of Global Regression *
***********************************************************************
Call:
lm(formula = formula, data = data)
Residuals:
Min 1Q Median 3Q Max
-3470778 -298119 -23481 248917 12234210
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 527633.22 108183.22 4.877 1.20e-06 ***
AREA_SQM 12777.52 367.48 34.771 < 2e-16 ***
AGE -24687.74 2754.84 -8.962 < 2e-16 ***
PROX_CBD -77131.32 5763.12 -13.384 < 2e-16 ***
PROX_CHILDCARE -318472.75 107959.51 -2.950 0.003231 **
PROX_ELDERLYCARE 185575.62 39901.86 4.651 3.61e-06 ***
PROX_URA_GROWTH_AREA 39163.25 11754.83 3.332 0.000885 ***
PROX_MRT -294745.11 56916.37 -5.179 2.56e-07 ***
PROX_PARK 570504.81 65507.03 8.709 < 2e-16 ***
PROX_PRIMARY_SCH 159856.14 60234.60 2.654 0.008046 **
PROX_SHOPPING_MALL -220947.25 36561.83 -6.043 1.93e-09 ***
PROX_BUS_STOP 682482.22 134513.24 5.074 4.42e-07 ***
NO_Of_UNITS -245.48 87.95 -2.791 0.005321 **
FAMILY_FRIENDLY 146307.58 46893.02 3.120 0.001845 **
FREEHOLD 350599.81 48506.48 7.228 7.98e-13 ***
---Significance stars
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 756000 on 1421 degrees of freedom
Multiple R-squared: 0.6507
Adjusted R-squared: 0.6472
F-statistic: 189.1 on 14 and 1421 DF, p-value: < 2.2e-16
***Extra Diagnostic information
Residual sum of squares: 8.120609e+14
Sigma(hat): 752522.9
AIC: 42966.76
AICc: 42967.14
BIC: 41731.39
***********************************************************************
* Results of Geographically Weighted Regression *
***********************************************************************
*********************Model calibration information*********************
Kernel function: gaussian
Adaptive bandwidth: 30 (number of nearest neighbours)
Regression points: the same locations as observations are used.
Distance metric: Euclidean distance metric is used.
****************Summary of GWR coefficient estimates:******************
Min. 1st Qu. Median 3rd Qu.
Intercept -1.3487e+08 -2.4669e+05 7.7928e+05 1.6194e+06
AREA_SQM 3.3188e+03 5.6285e+03 7.7825e+03 1.2738e+04
AGE -9.6746e+04 -2.9288e+04 -1.4043e+04 -5.6119e+03
PROX_CBD -2.5330e+06 -1.6256e+05 -7.7242e+04 2.6624e+03
PROX_CHILDCARE -1.2790e+06 -2.0175e+05 8.7158e+03 3.7778e+05
PROX_ELDERLYCARE -1.6212e+06 -9.2050e+04 6.1029e+04 2.8184e+05
PROX_URA_GROWTH_AREA -7.2686e+06 -3.0350e+04 4.5869e+04 2.4613e+05
PROX_MRT -4.3781e+07 -6.7282e+05 -2.2115e+05 -7.4593e+04
PROX_PARK -2.9020e+06 -1.6782e+05 1.1601e+05 4.6572e+05
PROX_PRIMARY_SCH -8.6418e+05 -1.6627e+05 -7.7853e+03 4.3222e+05
PROX_SHOPPING_MALL -1.8272e+06 -1.3175e+05 -1.4049e+04 1.3799e+05
PROX_BUS_STOP -2.0579e+06 -7.1461e+04 4.1104e+05 1.2071e+06
NO_Of_UNITS -2.1993e+03 -2.3685e+02 -3.4699e+01 1.1657e+02
FAMILY_FRIENDLY -5.9879e+05 -5.0927e+04 2.6173e+04 2.2481e+05
FREEHOLD -1.6340e+05 4.0765e+04 1.9023e+05 3.7960e+05
Max.
Intercept 18758355
AREA_SQM 23064
AGE 13303
PROX_CBD 11346650
PROX_CHILDCARE 2892127
PROX_ELDERLYCARE 2465671
PROX_URA_GROWTH_AREA 7384059
PROX_MRT 1186242
PROX_PARK 2588497
PROX_PRIMARY_SCH 3381462
PROX_SHOPPING_MALL 38038564
PROX_BUS_STOP 12081592
NO_Of_UNITS 1010
FAMILY_FRIENDLY 2072414
FREEHOLD 1813995
************************Diagnostic information*************************
Number of data points: 1436
Effective number of parameters (2trace(S) - trace(S'S)): 350.3088
Effective degrees of freedom (n-2trace(S) + trace(S'S)): 1085.691
AICc (GWR book, Fotheringham, et al. 2002, p. 61, eq 2.33): 41982.22
AIC (GWR book, Fotheringham, et al. 2002,GWR p. 96, eq. 4.22): 41546.74
BIC (GWR book, Fotheringham, et al. 2002,GWR p. 61, eq. 2.34): 41914.08
Residual sum of squares: 2.528227e+14
R-square value: 0.8912425
Adjusted R-square value: 0.8561185
***********************************************************************
Program stops at: 2024-10-16 23:56:56.247667
4.3 Visualising SDF Fields
To visualise the fields in SDF, we need to first covert it into sf data.frame by using the code chunk below.
gwr_adaptive_output <- as.data.frame(gwr_adaptive$SDF) %>%
select(-c(2:15))
gwr_sf_adaptive <- cbind(condo_resale_sf,
gwr_adaptive_output)
glimpse(gwr_sf_adaptive)Rows: 1,436
Columns: 63
$ nb <nb> <66, 77, 123, 238, 239, 343>, <21, 162, 163, 19…
$ wt <list> <0.1666667, 0.1666667, 0.1666667, 0.1666667, …
$ POSTCODE <dbl> 118635, 288420, 267833, 258380, 467169, 466472…
$ SELLING_PRICE <dbl> 3000000, 3880000, 3325000, 4250000, 1400000, 1…
$ AREA_SQM <dbl> 309, 290, 248, 127, 145, 139, 218, 141, 165, 1…
$ AGE <dbl> 30, 32, 33, 7, 28, 22, 24, 24, 27, 31, 17, 22,…
$ PROX_CBD <dbl> 7.941259, 6.609797, 6.898000, 4.038861, 11.783…
$ PROX_CHILDCARE <dbl> 0.16597932, 0.28027246, 0.42922669, 0.39473543…
$ PROX_ELDERLYCARE <dbl> 2.5198118, 1.9333338, 0.5021395, 1.9910316, 1.…
$ PROX_URA_GROWTH_AREA <dbl> 6.618741, 7.505109, 6.463887, 4.906512, 6.4106…
$ PROX_HAWKER_MARKET <dbl> 1.76542207, 0.54507614, 0.37789301, 1.68259969…
$ PROX_KINDERGARTEN <dbl> 0.05835552, 0.61592412, 0.14120309, 0.38200076…
$ PROX_MRT <dbl> 0.5607188, 0.6584461, 0.3053433, 0.6910183, 0.…
$ PROX_PARK <dbl> 1.1710446, 0.1992269, 0.2779886, 0.9832843, 0.…
$ PROX_PRIMARY_SCH <dbl> 1.6340256, 0.9747834, 1.4715016, 1.4546324, 0.…
$ PROX_TOP_PRIMARY_SCH <dbl> 3.3273195, 0.9747834, 1.4715016, 2.3006394, 0.…
$ PROX_SHOPPING_MALL <dbl> 2.2102717, 2.9374279, 1.2256850, 0.3525671, 1.…
$ PROX_SUPERMARKET <dbl> 0.9103958, 0.5900617, 0.4135583, 0.4162219, 0.…
$ PROX_BUS_STOP <dbl> 0.10336166, 0.28673408, 0.28504777, 0.29872340…
$ NO_Of_UNITS <dbl> 18, 20, 27, 30, 30, 31, 32, 32, 32, 32, 34, 34…
$ FAMILY_FRIENDLY <dbl> 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0…
$ FREEHOLD <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1…
$ LEASEHOLD_99YR <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ MLR_RES <dbl> -1489099.55, 415494.57, 194129.69, 1088992.71,…
$ Intercept <dbl> 2050011.67, 1633128.24, 3433608.17, 234358.91,…
$ y <dbl> 3000000, 3880000, 3325000, 4250000, 1400000, 1…
$ yhat <dbl> 2886531.8, 3466801.5, 3616527.2, 5435481.6, 13…
$ residual <dbl> 113468.16, 413198.52, -291527.20, -1185481.63,…
$ CV_Score <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ Stud_residual <dbl> 0.38207013, 1.01433140, -0.83780678, -2.846146…
$ Intercept_SE <dbl> 516105.5, 488083.5, 963711.4, 444185.5, 211962…
$ AREA_SQM_SE <dbl> 823.2860, 825.2380, 988.2240, 617.4007, 1376.2…
$ AGE_SE <dbl> 5889.782, 6226.916, 6510.236, 6010.511, 8180.3…
$ PROX_CBD_SE <dbl> 37411.22, 23615.06, 56103.77, 469337.41, 41064…
$ PROX_CHILDCARE_SE <dbl> 319111.1, 299705.3, 349128.5, 304965.2, 698720…
$ PROX_ELDERLYCARE_SE <dbl> 120633.34, 84546.69, 129687.07, 127150.69, 327…
$ PROX_URA_GROWTH_AREA_SE <dbl> 56207.39, 76956.50, 95774.60, 470762.12, 47433…
$ PROX_MRT_SE <dbl> 185181.3, 281133.9, 275483.7, 279877.1, 363830…
$ PROX_PARK_SE <dbl> 205499.6, 229358.7, 314124.3, 227249.4, 364580…
$ PROX_PRIMARY_SCH_SE <dbl> 152400.7, 165150.7, 196662.6, 240878.9, 249087…
$ PROX_SHOPPING_MALL_SE <dbl> 109268.8, 98906.8, 119913.3, 177104.1, 301032.…
$ PROX_BUS_STOP_SE <dbl> 600668.6, 410222.1, 464156.7, 562810.8, 740922…
$ NO_Of_UNITS_SE <dbl> 218.1258, 208.9410, 210.9828, 361.7767, 299.50…
$ FAMILY_FRIENDLY_SE <dbl> 131474.73, 114989.07, 146607.22, 108726.62, 16…
$ FREEHOLD_SE <dbl> 115954.0, 130110.0, 141031.5, 138239.1, 210641…
$ Intercept_TV <dbl> 3.9720784, 3.3460017, 3.5629010, 0.5276150, 1.…
$ AREA_SQM_TV <dbl> 11.614302, 20.087361, 13.247868, 33.577223, 4.…
$ AGE_TV <dbl> -1.6154474, -9.3441881, -4.1023685, -15.524301…
$ PROX_CBD_TV <dbl> -3.22582173, -6.32792021, -4.62353528, 5.17080…
$ PROX_CHILDCARE_TV <dbl> 1.000488185, 1.471786337, -0.344047555, 1.5766…
$ PROX_ELDERLYCARE_TV <dbl> -3.26126929, 3.84626245, 4.13191383, 2.4756745…
$ PROX_URA_GROWTH_AREA_TV <dbl> -2.846248368, -1.848971738, -2.648105057, -5.6…
$ PROX_MRT_TV <dbl> -1.61864578, -8.92998600, -3.40075727, -7.2870…
$ PROX_PARK_TV <dbl> -0.83749312, 2.28192684, 0.66565951, -3.340617…
$ PROX_PRIMARY_SCH_TV <dbl> 1.59230221, 6.70194543, 2.90580089, 12.9836104…
$ PROX_SHOPPING_MALL_TV <dbl> 2.753588422, -0.886626400, -1.056869486, -0.16…
$ PROX_BUS_STOP_TV <dbl> 2.0154464, 4.4941192, 3.0419145, 12.8383775, 0…
$ NO_Of_UNITS_TV <dbl> 0.480589953, -1.380026395, -0.045279967, -0.44…
$ FAMILY_FRIENDLY_TV <dbl> -0.06902748, 2.69655779, 0.04058290, 14.312764…
$ FREEHOLD_TV <dbl> 2.6213469, 3.0452799, 1.1970499, 8.7711485, 1.…
$ Local_R2 <dbl> 0.8846744, 0.8899773, 0.8947007, 0.9073605, 0.…
$ geometry <POINT [m]> POINT (22085.12 29951.54), POINT (25656.…
$ geometry.1 <POINT [m]> POINT (22085.12 29951.54), POINT (25656.…
The code chunks below is used to create an interactive point symbol map of local R2
tmap_mode("view")
tm_shape(mpsz)+
tm_polygons(alpha = 0.1) +
tm_shape(gwr_sf_adaptive) +
tm_dots(col = "Local_R2",
border.col = "gray60",
border.lwd = 1) +
tm_view(set.zoom.limits = c(11,14))tmap_mode("plot")tmap_options(check.and.fix = TRUE)
tmap_mode("view")
AREA_SQM_SE <- tm_shape(mpsz)+
tm_polygons(alpha = 0.1) +
tm_shape(gwr_sf_adaptive) +
tm_dots(col = "AREA_SQM_SE",
border.col = "gray60",
border.lwd = 1) +
tm_view(set.zoom.limits = c(11,14))
AREA_SQM_TV <- tm_shape(mpsz)+
tm_polygons(alpha = 0.1) +
tm_shape(gwr_sf_adaptive) +
tm_dots(col = "AREA_SQM_TV",
border.col = "gray60",
border.lwd = 1) +
tm_view(set.zoom.limits = c(11,14))
tmap_arrange(AREA_SQM_SE, AREA_SQM_TV,
asp=1, ncol=2,
sync = TRUE)tm_shape(mpsz[mpsz$REGION_N=="CENTRAL REGION", ])+
tm_polygons()+
tm_shape(gwr_sf_adaptive) +
tm_bubbles(col = "Local_R2",
size = 0.15,
border.col = "gray60",
border.lwd = 1)